void Init() {
this.state = State.Initial;
this.stack = new HWStack(10);
this.node = Push(null, XmlNodeType.Document, null);
this.node.IsEmpty = false;
this.sb = new StringBuilder();
this.name = new StringBuilder();
this.poptodepth = 0;
this.current = null;
this.partial = '\0';
this.endTag = null;
this.a = null;
this.apos = 0;
this.newnode = null;
this.rootCount = 0;
this.foundRoot = false;
}
void ValidateContent(Node node) {
if (this.dtd != null) {
// See if this element is allowed inside the current element.
// If it isn't, then auto-close elements until we find one
// that it is allowed to be in.
string name = this.nametable.Add(node.Name.ToUpper()); // DTD is in upper case
int i = 0;
int top = this.stack.Count-2;
if (node.DtdType != null) {
// it is a known element, let's see if it's allowed in the
// current context.
for (i = top; i>0; i--) {
Node n = (Node)this.stack[i];
if (n.IsEmpty)
continue; // we'll have to pop this one
ElementDecl f = n.DtdType;
if (f != null) {
if (f.Name == this.dtd.Name)
break; // can't pop the root element.
if (f.CanContain(name, this.dtd)) {
break;
}
else if (!f.EndTagOptional) {
// If the end tag is not optional then we can't
// auto-close it. We'll just have to live with the
// junk we've found and move on.
break;
}
}
else {
// Since we don't understand this tag anyway,
// we might as well allow this content!
break;
}
}
}
if (i == 0) {
// Tag was not found or is not allowed anywhere, ignore it and
// continue on.
}
else if (i < top) {
Node n = (Node)this.stack[top];
if (i == top - 1 && name == n.Name) {
// e.g. p not allowed inside p, not an interesting error.
} else {
string closing = "";
for (int k = top; k >= i+1; k--) {
if (closing != "") closing += ",";
Node n2 = (Node)this.stack[k];
closing += "<"+n2.Name+">";
}
Log("Element '{0}' not allowed inside '{1}', closing {2}.",
name, n.Name, closing);
}
this.state = State.AutoClose;
this.newnode = node;
Pop(); // save this new node until we pop the others
this.poptodepth = i+1;
}
}
}
public void CopyAttributes(Node n) {
for (int i = 0, len = n.attributes.Count; i < len; i++) {
Attribute a = (Attribute)n.attributes[i];
Attribute na = this.AddAttribute(a.Name, a.Value, a.QuoteChar, false);
na.DtdType = a.DtdType;
}
}
void Validate(Node node) {
if (this.dtd != null) {
ElementDecl e = this.dtd.FindElement(node.Name);
if (e != null) {
node.DtdType = e;
if (e.ContentModel.DeclaredContent == DeclaredContent.EMPTY)
node.IsEmpty = true;
}
}
}
void ValidateAttribute(Node node, Attribute a) {
ElementDecl e = node.DtdType;
if (e != null) {
AttDef ad = e.FindAttribute(a.Name);
if (ad != null) {
a.DtdType = ad;
}
}
}
void Pop() {
if (this.stack.Count > 1) {
this.node = (Node)this.stack.Pop();
}
}
void Pop()
{
if (_depth > 1) {
_depth--;
_node = _stack[_depth-1];
}
}
private bool ParseEndTag()
{
this.m_state = State.EndTag;
this.m_current.ReadChar(); // consume '/' char.
string name = this.ScanName(SgmlReader.tagterm);
char ch = this.m_current.SkipWhitespace();
if (ch != '>')
{
Log("Expected empty start tag '/>' sequence instead of '{0}'", ch);
this.m_current.ScanToEnd(null, "Recovering", ">");
}
this.m_current.ReadChar(); // consume '>'
this.m_endTag = name;
// Make sure there's a matching start tag for it.
bool caseInsensitive = (this.m_folding == CaseFolding.None);
this.m_node = (Node)this.m_stack[this.m_stack.Count - 1];
for (int i = this.m_stack.Count - 1; i > 0; i--)
{
Node n = (Node)this.m_stack[i];
if (string.Equals(n.Name, name, caseInsensitive ? StringComparison.OrdinalIgnoreCase : StringComparison.Ordinal))
{
this.m_endTag = n.Name;
return true;
}
}
Log("No matching start tag for '</{0}>'", name);
this.m_state = State.Markup;
return false;
}
private void Pop()
{
if (this.m_stack.Count > 1)
{
this.m_node = (Node)this.m_stack.Pop();
}
}
void Init()
{
_state = State.Initial;
_stack = new Node[10];
_size = 10;
_depth = 0;
_node = Push(null, XmlNodeType.Document, null);
_node.IsEmpty = false;
_sb = new StringBuilder();
_name = new StringBuilder();
_poptodepth = 0;
_current = null;
_partial = '\0';
_endTag = null;
_a = null;
_apos = 0;
_newnode = null;
_poptodepth = 0;
_rootCount = 0;
}
private void Init()
{
this.m_state = State.Initial;
this.m_stack = new HWStack(10);
this.m_node = Push(null, XmlNodeType.Document, null);
this.m_node.IsEmpty = false;
this.m_sb = new StringBuilder();
this.m_name = new StringBuilder();
this.m_poptodepth = 0;
this.m_current = null;
this.m_partial = '\0';
this.m_endTag = null;
this.m_a = null;
this.m_apos = 0;
this.m_newnode = null;
this.m_rootCount = 0;
this.m_foundRoot = false;
this.unknownNamespaces.Clear();
}
void Grow()
{
int inc = 10;
int newsize = _size+inc;
Node[] narray = new Node[newsize];
Array.Copy(_stack, narray, _size);
_size = newsize;
_stack = narray;
}
public void CopyAttributes(Node n)
{
for (int i = 0; i < n._attcount; i++) {
Attribute a = n._attributes[i];
Attribute na = this.AddAttribute(a.Name, a.Value, a.QuoteChar);
na.DtdType = a.DtdType;
}
}
Node Push(string name, XmlNodeType nt, string value)
{
if (_depth == _size) Grow();
Node result;
if (_stack[_depth] == null) {
result = new Node(name, nt, value);
_stack[_depth] = result;
}
else {
result = _stack[_depth];
result.Reset(name, nt, value);
}
_depth++;
_node = result;
return result;
}
Node Push(string name, XmlNodeType nt, string value) {
Node result = (Node)this.stack.Push();
if (result == null) {
result = new Node();
this.stack[this.stack.Count-1] = result;
}
result.Reset(name, nt, value);
this.node = result;
return result;
}
private void ValidateContent(Node node)
{
if (node.NodeType == XmlNodeType.Element)
{
if (!VerifyName(node.Name))
{
Pop();
Push(null, XmlNodeType.Text, "<" + node.Name + ">");
return;
}
}
if (this.m_dtd != null)
{
// See if this element is allowed inside the current element.
// If it isn't, then auto-close elements until we find one
// that it is allowed to be in.
string name = node.Name.ToUpperInvariant(); // DTD is in upper case
int i = 0;
int top = this.m_stack.Count - 2;
if (node.DtdType != null) {
// it is a known element, let's see if it's allowed in the
// current context.
for (i = top; i > 0; i--)
{
Node n = (Node)this.m_stack[i];
if (n.IsEmpty)
continue; // we'll have to pop this one
ElementDecl f = n.DtdType;
if (f != null)
{
if ((i == 2) && string.Equals(f.Name, "BODY", StringComparison.OrdinalIgnoreCase)) // NOTE (steveb): never close the BODY tag too early
break;
else if (string.Equals(f.Name, this.m_dtd.Name, StringComparison.OrdinalIgnoreCase))
break; // can't pop the root element.
else if (f.CanContain(name, this.m_dtd))
{
break;
}
else if (!f.EndTagOptional)
{
// If the end tag is not optional then we can't
// auto-close it. We'll just have to live with the
// junk we've found and move on.
break;
}
}
else
{
// Since we don't understand this tag anyway,
// we might as well allow this content!
break;
}
}
}
if (i == 0)
{
// Tag was not found or is not allowed anywhere, ignore it and
// continue on.
return;
}
else if (i < top)
{
Node n = (Node)this.m_stack[top];
if (i == top - 1 && string.Equals(name, n.Name, StringComparison.OrdinalIgnoreCase))
{
// e.g. p not allowed inside p, not an interesting error.
}
else
{
#if DEBUG
string closing = "";
for (int k = top; k >= i+1; k--) {
if (closing != "") closing += ",";
Node n2 = (Node)this.m_stack[k];
closing += "<" + n2.Name + ">";
}
Log("Element '{0}' not allowed inside '{1}', closing {2}.", name, n.Name, closing);
#endif
}
this.m_state = State.AutoClose;
this.m_newnode = node;
Pop(); // save this new node until we pop the others
this.m_poptodepth = i + 1;
}
}
}
Node Push(Node n) {
// we have to do a deep clone of the Node object because
// it is reused in the stack.
Node n2 = Push(n.Name, n.NodeType, n.Value);
n2.DtdType = n.DtdType;
n2.IsEmpty = n.IsEmpty;
n2.Space = n.Space;
n2.XmlLang = n.XmlLang;
n2.CurrentState = n.CurrentState;
n2.CopyAttributes(n);
this.node = n2;
return n2;
}
bool ParseEndTag() {
this.state = State.EndTag;
this.current.ReadChar(); // consume '/' char.
string name = this.ScanName(SgmlReader.tagterm);
char ch = this.current.SkipWhitespace();
if (ch != '>') {
Log("Expected empty start tag '/>' sequence instead of '{0}'", ch);
this.current.ScanToEnd(null, "Recovering", ">");
}
this.current.ReadChar(); // consume '>'
this.endTag = name;
// Make sure there's a matching start tag for it.
bool caseInsensitive = (this.folding == CaseFolding.None);
this.node = (Node)this.stack[this.stack.Count-1];
for (int i = this.stack.Count-1; i>0; i--) {
Node n = (Node)this.stack[i];
if (caseInsensitive && string.Compare(n.Name, name, true) == 0) {
this.endTag = n.Name;
return true;
} else if ((object)n.Name == (object)name) {
return true;
}
}
Log("No matching start tag for '</{0}>'", name);
this.state = State.Markup;
return false;
}
public override bool Read() {
if (current == null) {
OpenInput();
}
State start = this.state;
if (node.Simulated) {
// return the next node
node.Simulated = false;
this.node = Top();
this.state = this.node.CurrentState;
return true;
}
bool foundnode = false;
while (! foundnode) {
switch (this.state) {
case State.Initial:
this.state = State.Markup;
this.current.ReadChar();
goto case State.Markup;
case State.Eof:
if (this.current.Parent != null) {
this.current.Close();
this.current = this.current.Parent;
} else {
return false;
}
break;
case State.EndTag:
if (this.endTag == (object)this.node.Name) {
Pop(); // we're done!
this.state = State.Markup;
goto case State.Markup;
}
Pop(); // close one element
foundnode = true;// return another end element.
break;
case State.Markup:
if (this.node.IsEmpty) {
Pop();
}
Node n = this.node;
foundnode = ParseMarkup();
break;
case State.PartialTag:
Pop(); // remove text node.
this.state = State.Markup;
foundnode = ParseTag(this.partial);
break;
case State.PseudoStartTag:
foundnode = ParseStartTag('<');
break;
case State.AutoClose:
Pop(); // close next node.
if (this.stack.Count <= this.poptodepth) {
this.state = State.Markup;
if (this.newnode != null) {
Push(this.newnode); // now we're ready to start the new node.
this.newnode = null;
this.state = State.Markup;
} else if (this.node.NodeType == XmlNodeType.Document) {
this.state = State.Eof;
goto case State.Eof;
}
}
foundnode = true;
break;
case State.CData:
foundnode = ParseCData();
break;
case State.Attr:
goto case State.AttrValue;
case State.AttrValue:
this.state = State.Markup;
goto case State.Markup;
case State.Text:
Pop();
goto case State.Markup;
case State.PartialText:
if (ParseText(this.current.Lastchar, false)) {
this.node.NodeType = XmlNodeType.Whitespace;
}
foundnode = true;
break;
}
if (foundnode && this.node.NodeType == XmlNodeType.Whitespace && this.whitespaceHandling == WhitespaceHandling.None) {
// strip out whitespace (caller is probably pretty printing the XML).
foundnode = false;
}
if (!foundnode && this.state == State.Eof && this.stack.Count>1) {
this.poptodepth = 1;
state = State.AutoClose;
this.node = Top();
return true;
}
}
if (!foundRoot && (this.NodeType == XmlNodeType.Element ||
this.NodeType == XmlNodeType.Text ||
this.NodeType == XmlNodeType.CDATA)) {
foundRoot = true;
if (this.IsHtml && (this.NodeType != XmlNodeType.Element ||
string.Compare(this.LocalName, "html", StringComparison.InvariantCultureIgnoreCase) != 0)) {
// Simulate an HTML root element!
this.node.CurrentState = this.state;
Node root = Push("html", XmlNodeType.Element, null);
SwapTopNodes(); // make html the outer element.
this.node = root;
root.Simulated = true;
root.IsEmpty = false;
this.state = State.Markup;
//this.state = State.PseudoStartTag;
//this.startTag = name;
}
return true;
}
return true;
}
bool ParseEndTag()
{
_state = State.EndTag;
_current.ReadChar(); // consume '/' char.
string name = _current.ScanToken(_sb, _tagterm, false);
char ch = _current.SkipWhitespace();
if (ch != '>') {
Log("Expected empty start tag '/>' sequence instead of '{0}'", ch);
_current.ScanToEnd(null, "Recovering", ">");
}
_current.ReadChar(); // consume '>'
// Make sure there's a matching start tag for it.
_endTag = _nametable.Add(name.ToLower());
_node = _stack[_depth-1];
for (int i = _depth-1; i>0; i--) {
if ((object)_stack[i].Name == _endTag)
return true;
}
Log("No matching start tag for '</{0}>'", name);
_state = State.Markup;
return false;
}